Acute respiratory tract infections are one of the most common reasons for hospital admission in adults. In clinical practice antibiotics are nearly universally administered because a specific microbial diagnosis is often not made. Recent evidence using sensitive molecular techniques indicate that a significant proportion of these illnesses are due to viruses. Although viral infection may predispose to secondary bacterial infections, the actual incidence has been poorly studied and is likely overestimated. If antibiotic use was targeted to only those patients with true bacterial co-infection and not to those with viral infection alone, overall patient care would be improved by reducing antibiotic related complications, spread of antimicrobial resistant bacteria, and overall costs. In this proposal we intend to document the rate of bacterial in adults admitted to the hospital with confirmed viral infection. We hypothesize that it is possible to identify patients with documented viral infections who are at low risk for bacterial complications by using clinical and laboratory parameters in combination with new serum biomarkers such as pro-calcitonin. We also believe that physicians will respond appropriately, by withholding or discontinuing antibiotics, in most patients with documented viral illnesses if considered to be at low risk for bacterial complications. In years1-2 we propose to recruit and carefully evaluate ~1000 adults admitted to the hospital with respiratory tract infections during 2 winters for the presence of viral and bacterial infection using standard and new sensitive molecular techniques. We will define a set of clinical and laboratory variables that accurately predict patients with viral infection who are at low risk for bacterial co-infection. In years 3-5, we will prospectively enroll a similar patient population of ~1500 subjects, and identify subjects with documented viral respiratory tract infections and who meet previously defined criteria predicting a low risk of bacterial co-infection. These subjects will be randomized to one of two study arms in a randomized controlled intervention study. Half will enter an "intervention" arm in which physicians receive information regarding the presence of a viral infection and "low bacterial risk status" along with a recommendation to withhold or discontinue antibiotics. The other half will receive "standard care" in which antibiotic use is administered at the discretion of the attending physician who will not receive additional information regarding viral diagnosis or bacterial risk status. The primary analysis will determine if there is a significant reduction in antibiotic use in patients in the intervention group compared to those in the standard care group. Secondary analysis will examine whether the intervention group has improved clinical outcomes (shorter hospitalization, less antibiotic related adverse events, similar or less morbidity and mortality, similar or fewer readmissions for respiratory illness) than the standard care group. PUBLIC HEALTH RELEVANCE This project proposes to decrease unnecessary antibiotic use in patients hospitalized with documented viral respiratory infections. If successful, the major effects on public health would be to limit the spread of antimicrobial resistant bacteria and to decrease health care costs associated antibiotic use and complications in the treatment of respiratory tract infections.